• # ODF et OOXML

    Posté par . En réponse au journal ODF mort-né ?. Évalué à 4.

    Extrait d'un mail sur la liste gnome-foundation, qui compare les deux formats et justifie la décision de GNOME d'appointer quelqu'un à l'ECMA pour participer aux discussions sur OOXML. La réalité semble, étrangement, plus nuancée que la vision des fanboys qui monopolisent ces forums.

    http://mail.gnome.org/archives/foundation-list/2007-October/(...)

    « Legally I am on thin ice, the council I've contacted opined strictly
    on Gnumeric's ability to ship filters.

    Organizationally the ODF has a somewhat more open standardization
    process. In that the OO.o implementers are involved directly.
    Whereas the ECMA TC could mostly review only the documentation, and
    not make substantive changes to the format itself. On the other
    hand, OASIS has been less welcoming of non-profit membership than
    ECMA.

    Technically the case is far from clear cut. OOX and ODF each have
    significant flaws. Each of them abound with questionable design
    elements that limit 3rd party implementions, cause performance
    problems, and documentation that at times seems to cover only the
    simple cases, while ignoring the important elements. The FUD
    flowing through the community has played on our rational mistrust of
    Microsoft's intentions, to distract from the details.

    There are two truths that need to be accepted
    1) ODF is an excellent start for OO.o's file format, but it is not
    perfect and will never be 'the one true office format' for all
    office applications without destroying it's utility by diluting
    it with so much random cruft that no implementation would be
    complete, and interoperability would suffer.

    2) OOX is a file format that is in use, and we will have to interact
    with it. The opportunity to improve the spec and have MS answer
    questions and clarify necessary details should not be wasted.
    It's not as good as having the source code to OO.o there to read
    yourself (which is why free software will eventually dominate)
    but it is a step forward.

    How we proceed from there is open for debate. My contention is that
    the FLOSS community should do what it does best, write good software.
    We should support the use of free software and advocate for Gnumeric
    or AbiWord or OO.o on their plentiful merits. Interested community
    members should also join both ODF and OOX committees to help improve
    the specifications. The number of people actually involved in the
    standards is pitifully small considering the size of the problem
    space, and the centrality of 'office-ish' applications. »